search

rapidsai-community / notebooks-contrib

RAPIDS Community Notebooks
Apache License 2.0
513 stars 267 forks source link

[FEA] Notebook on Tabular Data + SHAP #328

Open erinkhoo opened 3 years ago

erinkhoo commented 3 years ago

Is your feature request related to a problem? Please describe. I can't find a great example of cuML on a tabular novel dataset using SHAP

Describe the solution you'd like I want to PR a notebook explaining a demo with a new GPUTree SHAP project Provide a cuML leveraged notebook on a large recruitment tabular dataset to show the speed and usefulness of GPUTree SHAP

Describe alternatives you've considered Reviewing existing notebooks and not sharing my work.

Additional context https://www.kaggle.com/c/jane-street-market-prediction/discussion/200632

taureandyernv commented 3 years ago

Hey @erinkhoo ! We'd love to include your work. Please do make a PR to the README and add it as a link or into community tutorials folder. I'll review.

taureandyernv commented 3 years ago

@erinkhoo the SHAP notebook looks cool, but requires the Kaggle kernel to run the Janestreet module, which is highly restricted. I really want to credit you with the work. May you either

  1. create the GPU version of your notebook in Kaggle, with instructions, and I'll link to that notebook
  2. create a stand alone SHAP notebook with less restricted data and PR it?

Please let us know!